Bug #36038 Test parts.partition_basic_<engine> failing on Windows
Submitted: 13 Apr 2008 18:58 Modified: 1 May 2008 19:16
Reporter: Patrick Crews Email Updates:
Status: Closed Impact on me:
None 
Category:Tests: Server Severity:S3 (Non-critical)
Version:5.1 OS:Windows
Assigned to: Patrick Crews CPU Architecture:Any
Tags: partition, parts, Tests, windows

[13 Apr 2008 18:58] Patrick Crews
Description:
Test parts.partition_basic_<engine> is failing on Windows with the following errors:

ls: cannot access f:/build/mysql-5.1.24-rc-winbuild/mysql-5.1.24-rc-win32/mysql-test/var/master-data/test/t1*: No such file or directory (happened on pushbuild)

Was unable to duplicate this on my own, but came up with these failures (see also Bug#36036, Bug#36037)

Example output:
@@ -383,16 +383,16 @@
22738:  WHERE f_int1 BETWEEN @max_row_div2 AND @max_row;
22739:  # Start usability test (inc/partition_check.inc)
22740:  create_command
22741: -SHOW CREATE TABLE t1;
22742: -Table	Create Table
22743: -t1	CREATE TABLE `t1` (
22744: -  `f_int1` int(11) NOT NULL DEFAULT '0',
22745: -  `f_int2` int(11) NOT NULL DEFAULT '0',
22746: -  `f_char1` char(20) DEFAULT NULL,
22747: -  `f_char2` char(20) DEFAULT NULL,
22748: -  `f_charbig` varchar(1000) DEFAULT NULL,
22749: -  PRIMARY KEY (`f_int1`,`f_int2`)
22750: -) ENGINE=InnoDB DEFAULT CHARSET=latin1 /*!50100 PARTITION BY HASH (f_int1)
PARTITIONS 2  */
22751: +SHOW CREATE TABLE t1;
22752: +Table	Create Table
22753: +t1	CREATE TABLE `t1` (
22754: +  `f_int1` int(11) NOT NULL DEFAULT '0',
22755: +  `f_int2` int(11) NOT NULL DEFAULT '0',
22756: +  `f_char1` char(20) DEFAULT NULL,
22757: +  `f_char2` char(20) DEFAULT NULL,
22758: +  `f_charbig` varchar(1000) DEFAULT NULL,
22759: +  PRIMARY KEY (`f_int1`,`f_int2`)
22760: +) ENGINE=InnoDB DEFAULT CHARSET=latin1 /*!50100 PARTITION BY HASH (f_int1)
PARTITIONS 2  */

How to repeat:
Run the test on a Windows machine.

From the mysql-test directory, execute the following command:

./mysql-test-run.pl --suite=parts partition_basic_myisam

You will see similar test failures.

Suggested fix:
Need to determine what is causing the perceived difference in results (expected vs.
actual).

Will disable the offending subtests in the meantime to maintain some test functionality
while the problem is being addressed.
[1 May 2008 6:16] Bugs System
Pushed into 5.1.25-rc
[1 May 2008 6:19] Bugs System
Pushed into 6.0.6-alpha
[1 May 2008 19:16] Paul DuBois
Test suite changes. No changelog entry needed.